Senior Software Engineer - AI & Agentic Architecture

Description

The company behind the PlayStation brand is a subsidiary of Sony Group Corporation. It makes hardware and network services used by more than 100 million people and also develops entertainment, including well-known intellectual properties.

This is a senior individual-contributor role. You will write code hands-on while working with several agentic frameworks and SDKs, harnesses, MCP, and APIs. You will also provide technical leadership and work closely with engineering, security, and product teams. The work is building agentic systems, developer tooling, and reusable platform capabilities that make people across the company more productive.

Responsibilities

• Build and run production-quality agentic applications with agentic SDKs and harnesses. This covers multi-agent orchestration, delegation to subagents, tool integration, and lifecycle hooks.
• Design Model Context Protocol (MCP) tool and resource interfaces so coding agents and harnesses can reach backend systems and enterprise data safely and reliably.
• Write prompts that produce dependable structured output using JSON schemas, few-shot examples, and validation with retries. Build evaluation frameworks to measure and raise output quality.
• Manage context across long documents, multi-turn conversations, and handoffs between agents. Design escalation paths and human-in-the-loop steps that keep production systems reliable.
• Set up Codex and Claude Code for use across the team. This includes AGENTS.md and CLAUDE.md hierarchies, custom slash commands, Agent Skills, path-specific rules, and plan mode.
• Add agentic development to CI/CD pipelines for automated code review, test generation, and pull request feedback. Tune it to cut false positives and duplicate findings.
• Build and maintain project-scoped and user-scoped MCP server integrations that extend agent-native development across engineering workflows.
• Find the best places to apply AI-assisted development. Build reusable capabilities that several teams can adopt, rather than one-off fixes.
• Work directly with engineering teams through hands-on discovery to learn their workflows, pain points, and opportunities.
• Prototype and test new AI capabilities quickly with real users. Move the ones that work into secure, observable, maintainable production systems.
• Own the full lifecycle of your work: implementation, testing, deployment, monitoring, and iteration.
• Lead technically through architecture discussions, code reviews, and troubleshooting of agentic systems.
• Mentor engineers on agentic design patterns, prompt engineering, and agent-native development practices.
• Explain tradeoffs, risks, and results clearly to technical and non-technical stakeholders.
• Contribute engineering standards, reusable patterns, and documentation that help other teams use agentic development tools well.

Requirements

• You have 6+ years of professional software engineering experience, including substantial hands-on work building and running AI/LLM-powered or agentic applications.
• You write strong code in Java, Python, Go, Rust, or C++ and have solid fundamentals in API design, testing, debugging, code review, and maintainable code.
• You have practical production experience with agentic SDKs and harnesses, MCP, and LLMs, including tool use, orchestration, RAG, context management, structured output, and evaluation.
• You have designed MCP tool and resource interfaces and can reason about tool selection, error handling, and reliability in multi-agent systems.
• You have shown experience bringing AI into CI/CD and developer workflows.
• You can work through unclear problems, weigh tradeoffs, and make practical engineering decisions.
• You communicate well in writing and speaking, and you have mentored and led people through influence.
• You hold a bachelor's degree or equivalent.
• Preferred Qualifications: You have demonstrated expertise in agentic architecture, MCP integration, Codex and Claude Code configuration, prompt engineering, and context management.
• Preferred Qualifications: You have built multi-agent research or customer support systems, or structured data extraction pipelines.
• Preferred Qualifications: You have experience with AWS, Docker, Kubernetes, Terraform, and GitOps supporting AI or agentic infrastructure.
• Preferred Qualifications: You are familiar with vector databases, embeddings, and AI evaluation frameworks.
• Preferred Qualifications: You have contributed to open-source projects involving AI, agentic systems, or developer tooling.
• Preferred Qualifications: You think in systems. You understand how models, agents, tools, context, APIs, data, and infrastructure fit together in production, and you can reason about dependencies and tradeoffs across the stack.
• Preferred Qualifications: You adapt quickly to new AI tools, models, and frameworks as the field changes.

Conditions

Base pay is set using market benchmarking data and the geographic location of the role. The estimated base pay range is $177,300 - $265,900 USD.

The base pay range may vary with the hybrid working policy. Individual pay depends on job-related factors such as knowledge, skills, experience, and location.

Background checks are run at the offer stage for all new hires, which may include criminal checks for some roles. This requires processing personal information.

Equal opportunity applies.

Benefits

• The role is eligible for a top-tier benefits package covering medical, dental, vision, a matching 401(k), paid time off, a wellness program, and employee discounts on Sony products.
• The role may also be eligible for a bonus package.